ORIGINAL PAPER

Transcriptional Control of Glutaredoxin GRXC9 Expression by aSalicylic Acid-Dependent and NPR1-IndependentPathway in Arabidopsis

Ariel Herrera-Vásquez & Loreto Carvallo & Francisca Blanco & Mariola Tobar &

Eva Villarroel-Candia & Jesús Vicente-Carbajosa & Paula Salinas & Loreto Holuigue

Published online: 14 August 2014# The Author(s) 2014. This article is published with open access at Springerlink.com

Abstract Salicylic acid (SA) is a key hormone that mediatesgene transcriptional reprogramming in the context of thedefense response to stress. GRXC9, coding for a CC-typeglutaredoxin from Arabidopsis, is an SA-responsive geneinduced early and transiently by an NPR1-independent path-way. Here, we address the mechanism involved in this SA-dependent pathway, using GRXC9 as a model gene. We firstestablished that GRXC9 expression is induced by UVB expo-sure through this pathway, validating its activation in a phys-iological stress condition. GRXC9 promoter analyses indicatethat SA controls gene transcription through two activatingsequence-1 (as-1)-like elements located in its proximal region.TGA2 and TGA3, but not TGA1, are constitutively bound tothis promoter region. Accordingly, the transient recruitment ofRNA polymerase II to the GRXC9 promoter, as well as thetransient accumulation of gene transcripts detected in SA-treated WT plants, was abolished in a knockout mutant forthe TGA class II factors. We conclude that constitutive bind-ing of TGA2 is essential for controlling GRXC9 expression,

while binding of TGA3 in a lesser extent contributes to thisregulation. Finally, overexpression of GRXC9 indicates thatthe GRXC9 protein negatively controls its own gene expres-sion, forming part of the complex bound to the as-1-contain-ing promoter region. These findings are integrated in a modelthat explains how SA controls transcription of GRXC9 in thecontext of the defense response to stress.

Keywords as-1-like element . GlutaredoxinGRXC9(GRX480) . NPR1-independent . Salicylic acid . TGAtranscription factors

Introduction

Salicylic acid (SA) is a key plant hormone involved instress defense responses against a wide range ofbiotrophic and hemibiotrophic pathogens and abioticstress conditions such as UV, high light radiation,ozone exposure, salinity, osmotic, and drought stress(Borsani et al. 2001; Garcion et al. 2008; Lee et al.2010; Mateo et al . 2006; Miura et al . 2013;Wildermuth et al. 2001; Nawrath and Metraux 1999;Ogawa et al. 2005). In response to stress, SA triggers aglobal transcriptional reprogramming in the infected/damaged tissues, as well as in the neighboring cells,orchestrating local and systemic defense responses(Vlot et al. 2009; Fu and Dong 2013).

Recent reports support evidence that SA interplays withredox signals, such as H2O2 and glutathione, in the modula-tion of the defense response (Foyer and Noctor 2011;Dubreuil-Maurizi and Poinssot 2012; Noshi et al. 2012; Hanet al. 2013). Although this interplay is being increasinglyrecognized (Fu and Dong 2013), the precise mechanisms thatgovern this relationship are still unknown. We previously

The present address of Dra. Francisca Blanco is Centro de BiotecnologiaVegetal, Facultad de Ciencias Biológicas, UniversidadAndrés Bello, Chile.

Electronic supplementary material The online version of this article(doi:10.1007/s11105-014-0782-5) contains supplementary material,which is available to authorized users.

A. Herrera-Vásquez : L. Carvallo : F. Blanco :M. Tobar :E. Villarroel-Candia : P. Salinas : L. Holuigue (*)Departamento de Genética Molecular y Microbiología, Facultad deCiencias Biológicas, Pontificia Universidad Católica de Chile,Alameda 340, Santiago, Chilee-mail: lholuigue@bio.puc.cl

J. Vicente-CarbajosaCentro de Biotecnología y Genómica de Plantas (UPM-INIA),Universidad Politécnica de Madrid, 28223 Pozuelo de Alarcón,Madrid, Spain

Plant Mol Biol Rep (2015) 33:624–637DOI 10.1007/s11105-014-0782-5

reported that a subset of the early SA-inducible genes (earlySAIGs) code for enzymes with glutathione (GSH)-dependentantioxidant and detoxifying activities, such as glutaredoxins(GRX) and glutathione S-transferases (GST) (Blanco et al.2005, 2009). Moreover, we showed that the expression ofGRXS13, one of the GRXs coded by early SAIGs, is criticalfor limiting basal and high light stress-induced reactive oxy-gen species (ROS) production and for regulation of theascorbate/dehydroascorbate (ASC/DHA) ratio after stress(Laporte et al. 2012). These data support the idea that thesegenes could be involved in the ROS-scavenging/antioxidantnetwork that contains the oxidative burst produced understress conditions.

Here, we studied GRXC9 (also known as GRX480), asecond GRX gene identified as an early SAIG (Blancoet al. 2009). The GRXS13 and GRXC9 genes code forglutaredoxins belonging to the plant-specific CC-type inArabidopsis (Ndamukong et al. 2007; La Camera et al.2011; Laporte et al. 2012). Glutaredoxins are smalldisulfide oxidoreductases that catalyze the reduction ofdisulfide bridges and protein–GSH adducts (S-glutathionylated proteins) using the reducing power ofGSH and NADPH (Rouhier et al. 2008). In this work,we specifically focus in unrevealing the transcriptionalcontrol mechanisms of GRXC9 by SA.

Transcriptional activation of the majority of the SAIGs,including the pathogenesis-related 1 gene (PR-1 gene, themost frequently used marker gene for SA signaling) is medi-ated by the master coactivator nonexpressor of pathogenesis-related genes 1 (NPR1) (Fu and Dong 2013; Dong 2004). SAcontrols the nuclear targeting and activity of the NPR1 pro-tein, via posttranslational modifications and degradation (Fuet al. 2012; Mou et al. 2003; Tada et al. 2008). Recently, theNPR1 paralogs NPR3 and NPR4 were identified as directreceptors of SA that regulate NPR1 degradation, controllingin this way the SA responses mediated by this coactivator (Fuet al. 2012). On the other hand, there is a second pathway thatleads to the early and transient activation of SAIGs, via aNPR1-independent mechanism, as we and other groups havepreviously reported (Lieberherr et al. 2003; Uquillas et al.2004; Blanco et al. 2005, 2009; Fode et al. 2008; Langlois-Meurinne et al. 2005; Shearer et al. 2012). The mechanism bywhich SA activates the expression of these early SAIGs is stillunknown. In this work, we assess this mechanism usingGRXC9 as a model for the SA-dependent and NPR1-independent pathway that controls defense gene expression.

Promoter analyses of early SAIGs show overrepre-sentation of a cis-acting element with high homology tothe activating sequence-1 (as-1) (Blanco et al. 2005,2009). The as-1 element consists of two adjacent vari-ants of the palindromic sequence TGAC/GTCA(TGACG box), separated by four base pairs (Elliset al. 1993; Krawczyk et al. 2002). The as-1 element

was first identified in the Cauliflower Mosaic Virus 35S(CaMV 35S) promoter, as an element that conferred basalexpression in root tips (Benfey et al. 1989). Subsequent stud-ies showed that the as-1 element from the CaMV 35S pro-moter responds early and transiently to SA (Qin et al. 1994), toxenobiotic compounds like the synthetic auxin 2,4-dichlorophenoxyacetic acid (2,4D) (Johnson et al. 2001),and to H2O2 and methyl viologen (Garreton et al. 2002).Accordingly, this particular array of two adjacent TGACGboxes has been found overrepresented, not only in earlySAIGs promoters (Blanco et al. 2005, 2009) but also in thepromoters of plant genes associated to chemical detoxificationprocess induced by xenobiotic chemicals like 2,4-D and oxi-dized lipids (oxilipins) (Fode et al. 2008; Köster et al. 2012;Johnson et al. 2001).

TGACG boxes are recognized by basic leucine zipper(bZIP) factors of the TGA family, which has 10 members inArabidopsis (Jakoby et al. 2002; Gatz 2012). Seven of thesefactors have been associated to the defense response (TGA1-TGA7), being class II (TGA2, TGA5, and TGA6) the mostrelevant for the SA pathway (Gatz 2012). In fact, involvementof TGA class II factors in the canonic pathway that controlsexpression of SA- and NPR1-dependent genes containingTGA motifs in their promoters has been extensively reportedusing the Arabidopsis PR-1 gene as a model (Lebel et al.1998; Kesarwani et al. 2007). In contrast, the pathway thatactivates SA-dependent and NPR1-independent genes con-taining the as-1-like element in their promoters has been farless explored.

In this work, we show evidence that GRXC9 expres-sion is activated by stress, such as UVB radiation, viaan SA-dependent and NPR1-independent pathway. Bio-informatics analysis of the GRXC9 promoter indicatesthe presence of two putative as-1-like elements in itsproximal region. We show that both elements are func-tionally relevant and required for the SA-mediated in-duction of the gene. Moreover, our data indicate thatTGA2 and TGA3, but not TGA1, are constitutivelybound to the GRXC9 promoter in vivo. We also con-firm the requirement of TGA class II factors for therecruitment of the RNA polymerase II (Pol II) enzymeto the basal promoter and for the transcriptional induc-tion of the gene upon SA stimuli. Analysis of thetransactivation as well as the homodimerization andheterodimerization ability of TGA class II and TGA3factors gives further insights into the mechanism ofinduction of antioxidant genes by the SA/TGA2-3/as-1-like pathway. Interestingly, we show evidence indi-cating that GRXC9 participates in the control of theexpression of its own gene, suggesting its involvementin the transient activation of early SAIGs. Physiologi-cal and mechanistic implications of these findings arediscussed.

Plant Mol Biol Rep (2015) 33:624–637 625

Results

GRXC9 Gene Expression Is Induced by UVB Exposure, viaan SA-Dependent and NPR1-Independent Mechanism

We previously reported that GRXC9 gene expression is earlyand transiently induced by exogenous treatment with SA, viaan NPR1-independent pathway (Blanco et al. 2009). Weconfirmed these results by quantitative reverse transcriptionPCR (RT-qPCR), showing that a significant increase inGRXC9 transcript levels occurs after 2.5 h of SA treatmentin both WT and npr1-1 mutant plants (Online Resource 1).Considering that these results are in disagreement with aprevious report that claims that GRXC9 expression inducedby SA is dependent on NPR1 (Ndamukong et al. 2007), weused the npr3-1 npr4-3 double mutant to further assess thispoint. NPR3 and NPR4 proteins are required to control NPR1degradation (Fu et al. 2012). The npr3-1 npr4-3mutant showsincreased basal levels of NPR1 protein and, as a consequence,higher basal expression levels of NPR1-dependent PR1, PR2,and PR5 genes than theWT plants (Fu et al. 2012; Zhang et al.2006). In the case ofGRXC9 expression, we detected an earlyand transient increase of transcript levels in the npr3 npr4double mutants (Online Resource 1), while the basal levelsremained unchanged when compared to WT or npr1-1 plants(Online Resource 1, insert). This evidence confirms thatGRXC9 gene expression induced by SA treatment does notdepend on NPR1 levels.

To further validate the SA-dependent and NPR1-independent expression ofGRXC9 under physiological stress,we used UVB radiation as a stress condition since SA hasbeen identified as a signaling molecule in this defense re-sponse (Surplus et al. 1998). To evaluate the GRXC9 geneexpression dependence on SA, we used sid2-2 mutant plantsthat are deficient in SA biosynthesis (Wildermuth et al. 2001).Then,GRXC9 transcript levels were measured by RT-qPCR inWT, npr1-1 and sid2-2 Arabidopsis plants exposed for 2.5 and24 h to UVB light. Mean values from 3 to 6 replicates areshown in Fig. 1, while data from a representative replicate areshown in Online Resource 2. We detected significant increaseinGRXC9 transcript levels after 24 h of stress exposure inWTand npr1-1 plants. This response was almost completelyabolished in the sid2-2 mutant plants (Fig. 1a, Online Re-source 2a). PR-1 gene expression after UVB treatment wasevaluated as a control for the SA- and NPR1-dependentpathway (Fig. 1b, Online Resource 2b).

Concerning the basal levels of GRXC9 and PR-1expression, we detected a reduction in npr1-1 andsid2-2 mutants compared to WT plants, but these dif-ferences were not statistically significant (Fig. 1a, b,Online Resource 1 and 2). Together, these results showthat GRXC9 is an SA-dependent and NPR1-independentstress responsive gene. We used the GRXC9 gene as a

model to inquire about the mechanism involved in thispathway.

Two as-1-Like Elements in the GRXC9 Promoter AreRequired and Sufficient for SA-Mediated TranscriptionalActivation

In silico analysis of the GRXC9 promoter sequence revealedthe presence of several putative SA-responsive elements in-cluding two as-1-like elements, several W boxes, and isolatedTGACG boxes (Online Resource 3). The proximal as-1-likeelement is located between −80 and −99 bp, and the distal oneis located between −114 and −133 bp upstream of the tran-scriptional start site (Fig. 2a and Online Resource 3). In orderto evaluate whether these elements mediate the SA-dependenttranscriptional activation of the gene, we generatedArabidopsis transgenic lines harboring different versions ofthe GRXC9 promoter fused to the GUS reporter gene. Theconstructs contained either the complete intergenic region

Fig. 1 GRXC9 expression levels in WT, sid2-2, and npr1-1 seedlingsupon UVB chronic exposure. Expression levels of GRXC9 (a) and PR1(b) genes in 15-day-old seedlings from WT (black bars), npr1-1 (graybars), and sid2-2 (white bars) backgrounds were exposed to UVB light.The transcript levels for each gene were quantified by RT-qPCR fromsamples collected after 0, 2.5, and 24 h of UVB exposure. The relativeexpression was calculated by normalizing the GRXC9 and PR1 transcriptlevels to that of the YLS8 gene and to the WT basal levels. Error barsrepresent the mean±standard error from 3 to 6 replicates. Data from arepresentative replicate is shown in Online Resource 2. Letters above thebars indicate significant differences based on unpaired t test (n≥3,p<0.05)

626 Plant Mol Biol Rep (2015) 33:624–637

(−1,849 to +26), considered as the fullGRXC9 promoter (pC9WT::GUS), or truncated versions of this sequence that includethe two as-1-like elements up to −168 bp (pC9-168::GUS),only the proximal as-1-like element up to −112 bp (pC9-112::GUS), or a minimal promoter up to −61 that includesthe putative TATA box (pC9-61::GUS) (Fig. 2a and OnlineResource 3).

We treated seedlings (from six to 13 independent ho-mozygous lines for each construct) with 0.5× MS ascontrol or with 0.5 mM SA for 2.5 h. We then quantifiedthe basal and SA-induced GUS activities in total proteinextracts (Online Resource 4). The responsiveness to SA ofeach construct was represented as the mean value of theGUS activity induction ratio (SA-induced/basal GUS

activities) (Fig. 2a). An average of sixfold increase inGUS activity after SA treatment was recorded in linesthat contain the full promoter (pC9 WT::GUS), whichindicates that GRXC9 gene expression is effectively acti-vated by SA at the transcriptional level. Surprisingly, avery small promoter region up to −168 retains an impor-tant part of the SA responsiveness of the GRXC9 fullpromoter (Fig. 2a and Online Resource 4). This regioncontains both as-1-like elements, the most proximalTGACG box and the most proximal W box, while lackingall the rest of the putative SA-responsive elements (OnlineResource 3). In contrast, lines expressing the pC9-112: :GUS and the pC9-61: :GUS constructs werecompletely insensitive to SA treatment (Fig. 2a and

Fig. 2 GRXC9 promoter analysis by GUS reporter activity assays inArabidopsis lines (a). The diagram represents the GRXC9 promoterconstructs used to generate the GUS reporter lines. The numbers on theleft side indicate the size of the promoter region cloned to drive GUSexpression, and the numbers above the first construct indicate the positionfrom the transcriptional start site of the as-1-like elements in the GRXC9promoter. pC9 WT::GUS complete intergenic region for GRXC9, pC9-168::GUS promoter region containing the two as-1-like elements, pC9-112::GUS region containing the proximal as-1-like element, and pC9-61::GUS sequence of the promoter region containing the putative TATAbox. To quantify GUS activity induced by SA in the reporter lines, 6 to 13homozygous lines per construct were selected. Fifteen-day-old seedlingsfrom each line were treated with SA 0.5 mM or 0.5× MS as a control (C)for 2.5 h. GUS activities were quantified in total protein extracts fromeach independent line and normalized with the total the amount ofproteins (Online Resource 4). The ratio between SA treatment and itsrespective control was calculated, and the graph represents the average ofGUS activity ratio obtained from the different lines. Error bars representthe standard error from three independent experiments. Letters indicate

significant differences based on unpaired t test (p<0.05). b Schematicrepresentation of genetic constructs containing site-directed mutations inthe as-1-like elements, in the context of the full GRXC9 promoter se-quence, and fused to GUS-coding region. pC9 WT::GUS (WT in thescheme) was used as a template to mutate the TGACG boxes (indicatedby black arrows) from the two as-1-like elements (highlighted in blackboxes). The mutated base pairs in distal as-1-like element (MD) andproximal as-1-like element (MP) are indicated in lowercase andhighlighted in gray boxes. c Four independent homozygous GUS reporterlines carrying the pC9 WT::GUS and the mutated version in the proximaland distal as-1-like elements, pC9 MP::GUS and pC9 MD::GUS, respec-tively (described in c), were used to quantify GUS activity. Fifteen-day-old seedlings were treated with SA 0.5mM (black bars) or 0.5×MS (graybars) as control for 2.5 h, total proteins were prepared, and GUS activitywas quantified and normalized with total protein concentration. Thegraph shows the mean value of three biological replicates for each line.Error bars represent the ±standard error. Data from a representativereplicate is shown in Online Resource 5

Plant Mol Biol Rep (2015) 33:624–637 627

Online Resource 4), suggesting that the loss of the distalas-1-like element, and/or the proximal W and TGACGboxes, is enough to abolish SA-responsiveness.

To further evaluate the importance of the two as-1-likeelements in the SA-responsiveness of the promoter, we gen-erated genetic constructs containing the fullGRXC9 promotercarrying point mutations in each as-1-like element (Fig. 2b).The nucleotides mutated in the as-1-like sequences werechosen considering the most conserved ones in the consensusas-1-like element detected in the cluster of genes induced bySA in a NPR1-independent manner (Blanco et al. 2009). Fourindependent Arabidopsis reporter lines were selected for eachconstruct, either having mutations in the proximal (pC9 MP::-GUS) or in the distal (pC9 MD::GUS) as-1-like elements(Fig. 2b). Seedlings from each line were treated with SA or0.5× MS, and the GUS activity was quantified. As shown inFig. 2c (see also Online Resource 5 for data from a represen-tative replicate), lines that carry mutations in any of the twoas-1 elements no longer respond to SA treatment.

These results indicate that the increase inGRXC9 transcriptlevels in response to SA is mainly due to transcriptionalactivation of the gene mediated by this hormone and that theloss of any of the as-1-like elements is enough to abolish thisactivation. Both elements are functional, essential, and suffi-cient for SA responsiveness of the GRXC9 promoter.

SA-Induced GRXC9 Expression Is Abolishedin tga2-1/tga5-1/tga6-1 Triple Mutants, While TGA2and TGA3 Are Constitutively Bound to the GRXC9 PromoterIn Vivo

It has been reported thatGRXC9 induction by SA treatment isabolished in the TGA class II triple mutant (tga2-1/tga5-1/tga6-1) (Blanco et al. 2009; Ndamukong et al. 2007). Never-theless, the possible participation of other members of theTGA family proteins, as well as the direct binding of TGAfactors to the GRXC9 promoter, has not been addressed. Withthis purpose in mind, we first analyzed the SA-induced ex-pression of GRXC9 by RT-qPCR in different tga mutantbackgrounds (Fig. 3). Considering that TGAs belonging toclass I (TGA1 and TGA4) and class II (TGA2, TGA5, andTGA6) show different degrees of redundancy (Zhang et al.2003; Kesarwani et al. 2007), we used the double and triplemutant, respectively. The redundancy of TGA class III (TGA3and TGA7) has not been demonstrated, and thus we analyzedthe single mutants for each gene (Kesarwani et al. 2007). Wecorrelated the expression data with in vivo binding assays ofTGA factors to theGRXC9 promoter after SA treatment usingchromatin immunoprecipitation (ChIP) assays (Fig. 4). ChIP-qPCR assays were performed inWT plants treated with SA or0.5× MS as control, using primers that amplify a 290-bpfragment (−212 to +78 region) that includes the basal promot-er and the as-1-like elements (Fig. 4a). Antibodies that

specifically recognize TGA1, TGA2, and TGA3, raisedagainst the divergent N terminal regions, were used (Lamand Lam 1995).

In the double mutant of class I TGAs (tga1-1/tga4-1),GRXC9 expression was early and transiently activatedreaching its peak 2.5 h after SA treatment, as in WT plants(Fig. 3). Accordingly, the in vivo binding of TGA1 to theGRXC9 promoter, evaluated by ChIP-qPCR, was not detectedeither under control conditions or after 2.5 h of SA treatment(Fig. 4b). These results indicate that class I TGAs are notinvolved in GRXC9 induction.

In contrast, GRXC9 induction by SA was significantlyreduced in the triple mutant of TGA class II (tga2-1/tga5-1/tga6-1), compared to WT plants (Fig. 3), although a slightincrease of the transcript after 2.5 h of SA was observed. Onthe other hand, the induction level by SAwas 33.4 % reducedin the tga3-1 mutant compared to WT plants, albeit thisdifference was not statistically significant (Fig. 3). To betterevaluate the importance of TGA3 in GRXC9 expression, wealso assayed the quadruple tga2-1/tga3-1/tga5-1/tga6-1 mu-tant background. Although the lack of TGA class II had astriking negative effect on SA-induced GRXC9 transcription,the slight increase in messenger RNA (mRNA) levels after2.5 h of SA treatment seen in the triple mutant is no longerobserved in the quadruple mutant (Fig. 3). These resultssupport that TGA class II, and to a lesser extent TGA3, isinvolved in GRXC9 induction by SA. Interestingly, mutationsin TGA class II and TGA3 factors also have an effect on thebasal levels of GRXC9 expression (Fig. 3, insert). ComparedtoWT plants, basalGRXC9 transcript levels are reduced in thesingle tga3-1 and in the triple tga2-1/tga5-1/tga6-1 mutants,and this difference is higher and only statistically significant inthe quadruple tga2-1/tga3-1/tga5-1/tga6-1 mutant (Fig. 3 in-sert). Supporting these expression results, ChIP-qPCR assaysshow that TGA2 and TGA3 are constitutively bound to theGRXC9 promoter, either in the presence or in the absence ofSA stimulus (Fig. 4c, d).

Surprisingly, the lack of TGA7 produces a significantincrease in GRXC9 induction by SA (Fig. 3), suggesting thatTGA7 can play a negative role in this control mechanism. Wedid not detect involvement of TGA factors in repressingGRXC9 expression under basal conditions (Fig. 3, insert), incontrast to what was previously reported for PR-1 expression(Kesarwani et al. 2007).

Together, these results indicate that constitutive binding ofthe TGA2 factor to the GRXC9 promoter is essential fortranscriptional activation mediated by SA. Even thoughTGA3 is also constitutively bound, its role is more importantin the basal than in the SA-induced GRXC9 expression.

Considering that the constitutive binding of TGA2 andTGA3 to the GRXC9 promoter detected in vivo does notcorrelate with the constitutive expression of the gene, wepropose that these TGA factors could bind either as homo or

628 Plant Mol Biol Rep (2015) 33:624–637

heterodimers, without being directly able to transactivate tran-scription. In order to evaluate the potential homodimerizationand heterodimerization ability of TGA class II and TGA3factors, as well as their transactivation activity, we performedyeast one and two hybrid assays. For this purpose, we clonedthe CDS of TGA factors in frame with the DNA-bindingdomain (BD) or the activation domain (AD) of the yeastGal4 factor. Interactions between TGA factors were evaluatedin yeast by qualitative assays of the β-galactosidase reportergene, whose expression is controlled by four copies of theGal4-responsive element. The transactivation ability of theTGA factors was evaluated in assays using the TGA factorsfused to the BD-Gal4 and the empty pDEST22 vector. Ourresults indicate that TGA2, TGA3, TGA5, and TGA6 cannottransactivate the β-galactosidase gene in a yeast system,reflected by the null activity of the reporter enzyme (OnlineResource 6, first lane). Furthermore, the two-hybrid assaysshow that TGA class II and TGA3 factors are able tohomodimerize and also heterodimerize among them (OnlineResource 6). The interaction between NPR1 and TGA2 pro-teins was used as a positive control in these assays (Fan andDong 2002).

These results support the idea that TGA2 and TGA3 bindto the GRXC9 promoter, most probably through the SA-responsive as-1-like elements, both as homodimers or hetero-dimers, without acting directly in transactivation.

SA Induces the Transient Recruitment of RNA Polymerase IIto the GRXC9 Promoter by a TGA Class II-DependentMechanism

We evaluated whether, independently of the constitutive bind-ing of TGA2 and TGA3 to theGRXC9 promoter, the transientincrease in GRXC9 transcript levels correlates with a transient

recruitment of Pol II to the GRXC9 promoter, induced by SA.In order to test this, Arabidopsis seedlings were treated with0.5 mM SA or 0.5× MS as a control, and ChIP assays wereperformed at different times. We used antibodies that recog-nize the N-terminal domain of Pol II from Arabidopsis and theset of primers previously described (Fig. 4a). Interestingly, wefound a good correlation between the increment in GRXC9mRNA levels (Fig. 3, Online Resource 1a) and the recruit-ment of Pol II to the GRXC9 promoter triggered by SA(Fig. 5a). Similarly, the recruitment of Pol II, as well asGRXC9 expression, was abolished in the tga2-1/tga5-1/tga6-1 triple mutant (Figs. 5b and 3).

Taken together, these results indicate that SA triggers thetransient recruitment of Pol II to the basal GRXC9 promoter,which explains the transient increase in GRXC9 transcriptlevels triggered by SA. On the other hand, the fact that theincrease inGRXC9 transcript levels and the Pol II recruitmentare impaired in the tga2-1/tga5-1/tga6-1 triple mutant sug-gests that the constitutive binding of TGA factors is requiredfor differential Pol II recruitment to the promoter.

Overexpression of GRXC9 Downregulates the Expressionof its Endogenous Gene

It has been shown that TGA2 and GRXC9 are able to interactin vivo and that this interaction may have a role in controllingthe expression of genes involved in the defense responsetriggered by jasmonic acid (Ndamukong et al. 2007). Howev-er, the significance of this interaction in the SA response hasnot been addressed. Our results indicate that TGA2 is a keyfactor in the transcriptional induction of GRXC9, and thus weevaluated whether the overexpression ofGRXC9 has an effectin its own transient SA-dependent transcriptional induction.

Fig. 3 GRXC9 gene expression in tga mutant backgrounds. Expressionanalysis of the GRXC9 gene was evaluated by RT-qPCR in 15-day-oldseedlings of WT, tga1-1tga4-1 (tga14), tga3-1 (tga3), tga7-1 (tga7),tga2-1tga5-1tga6-1 (tga256), and tga2-1tga3-1tga5-1tga6-1 (tga2356)genotypes, under basal conditions (insert) and after treatment with SA0.5 mM or 0.5× MS as a control for 2.5 and 24 h. The GRXC9 relative

expression was calculated by normalizing the expression level ofGRXC9with the expression level of the housekeeping gene Clathrin adaptorcomplex subunit and with the WT basal condition. Error bars representthe ±standard error from three biological replicates (*p<0.05, comparedto the WT genotype in a two-way ANOVA and Bonferroni post test)

Plant Mol Biol Rep (2015) 33:624–637 629

We produced WT transgenic lines harboring the CaMV35S promoter controlling the expression of the GRXC9-cod-ing region fused to the immunological cMyc tag. We chosetwo homozygous lines (L3 and L7) with high constitutiveexpression levels of GRXC9 (Fig. 6a). We then evaluated the

levels of the endogenous GRXC9 transcript by RT-qPCR inWT and overexpressor lines treated with SA or 0.5× MS ascontrol for 2.5 h. As shown in Fig. 6b, SA induction of theendogenous GRXC9 gene was significantly reduced in bothoverexpressor lines to less than 50 % of the SA induction

�Fig. 4 Analysis of TGAs binding to as-1-like elements of the GRXC9promoter by ChIP-qPCR assays. a Diagram of the GRXC9 promoterregion amplified in the ChIP-qPCR assays. The arrowheads indicatethe location of the primers used to quantify the DNA of the GRXC9promoter bound to TGAs by qPCR. Fifteen-day-old plants treated with0.5 mM SA (SA) or 0.5× MS as control (C) for 2.5 and 24 h were used toperform ChIP-qPCR assays. Antibodies raised against TGA1 (b), TGA2(c), and TGA3 (d) transcription factors (in black bars) or IgG as a control(white bars) were used for the ChIP assays. qPCR analyses to quantify theDNA recovered from the ChIP were performed using the primersdescribed in a. The values for the immunoprecipitated DNA samplesare expressed as fold enrichment with the specific antibody over anonspecific immunoprecipitation condition (IgG). Error bars represent±standard error of three biological replicates

Fig. 5 Recruitment of the RNA Pol II to the GRXC9 promoter region byChIP-qPCR assays. WT (a) and tga2-1 tga 5–1 tga 6–1 (b) plants treatedwith 0.5 mM SA (SA) or 0.5×MS as control (C) for 0, 2.5, and 24 h wereused to perform ChIP-qPCR assays. The ChIP assays were performedwith a polyclonal antibody raised against the Pol II (black bars) or with apurified IgG (white bars) as a control. The qPCR analyses to quantify theDNA recovered from the ChIP were performed using the primers de-scribed in Fig. 4a. The values for the immunoprecipitated DNA sampleswere expressed as fold enrichment with the specific antibody over anonspecific immunoprecipitation condition (IgG). Error bars represent±standard error of three biological replicates

630 Plant Mol Biol Rep (2015) 33:624–637

observed in WT plants. Although basal levels of GRXC9expression are reduced in both overexpressor lines comparedto WT, these differences were not statistically significant(Fig. 6b). As a control, we show that overexpression ofGRXC9 does not affect basal or SA-induced PR-1 gene ex-pression (Online Resource 7). These results suggest thatGRXC9 negatively regulates the expression of its own gene.

To further investigate the role of GRXC9 in the regulationof its own gene, we used ChIP-qPCR assays in theoverexpressor lines to evaluate whether the GRXC9-Mycprotein is associated with the TGA2/TGA3 complex boundto the as-1-like elements in the GRXC9 promoter. Interesting-ly, we detected that GRXC9-Myc protein effectively formspart of the protein complex bound to the promoter (Fig. 6c),strongly suggesting that GRXC9 regulates its own gene ex-pression through binding to TGA factors while they are boundto the DNA.

Discussion

In this study, we explore the mechanism of control ofgene expression via an SA-dependent and NPR1-independent pathway in Arabidopsis, using GRXC9 as amodel gene. First, we showed evidence for the inductionof GRXC9 expression through this pathway in responseto UVB stress, validating its activation in the context ofa defense response to stress (Fig. 1a). By assaying inplanta GRXC9-GUS reporter activity (Fig. 2a) andin vivo binding of Pol II to the GRXC9 promoter(Fig. 5a), we showed that the control of GRXC9 geneexpression by SA is exerted at the initiation of transcrip-tion. Accordingly, we established that TGA class II fac-tors (Figs. 3 and 5b), as well as the two as-1-like ele-ments located in the GRXC9 proximal promoter(Fig. 2c), are essential for the induction of GRXC9 ex-pression by SA. The constitutive binding of TGA2 andTGA3 factors to the GRXC9 promoter, as detected byin vivo ChIP assays (Fig. 4), indicates that the inducingeffect of SA is not due to an increase in binding of TGAfactors to the as-1-like elements. TGA class II and TGA3factors have the capacity to interact with each other, asdetected in yeast by two-hybrid assays (Online Resource6), suggesting that TGA2 and TGA3 can bind to theGRXC9 promoter as homodimers or heterodimers. Fur-thermore, TGA class II and TGA3 factors did not showtransactivation capacity in yeast one-hybrid assays (On-line Resource 6). Therefore, even though TGA class IIfactors are essential for recruiting Pol II to the GRXC9promoter (Fig. 5b), additional coregulators are requiredfo r t r an s a c t i v a t i on . F i n a l l y, we showed tha toverexpressed GRXC9-Myc protein binds to the TGA-

Fig. 6 Effect ofGRXC9 overexpression onGRXC9 gene expression andon binding to the GRXC9 promoter. The levels of total (a) and endoge-nous (b) GRXC9 transcripts were detected by RT-qPCR in 15-day-oldseedlings of WT and GRXC9 OE lines (L3 and L7) treated with SA0.5mMor 0.5×MS as a control for 2.5 h.GRXC9 relative expressionwascalculated by normalizing the expression level of GRXC9 with the ex-pression level of the housekeeping gene YLS8 and to the WT basalcondition. Error bars represent the ±standard error. Letters above thebars indicate significant differences based on unpaired t test (n=3,p<0.05). The binding of GRXC9-Myc protein to the endogenous pro-moter of GRXC9 was evaluated by ChIP-qPCR (c). WT plants and thetwo overexpressor lines (OXC9 L3 and OXC9 L7) were evaluated in theconditions mentioned above. The immunoprecipitation was performedwith commercial antibodies raised against the MYC-tag (black bars) anda nonspecific IgG (white bars) as control. The qPCR analyses to quantifythe DNA recovered from the ChIP were performed using the primersdescribed in Fig. 4a. The values for the immunoprecipitated DNA sam-ples were expressed as fold enrichment of the specific antibody over anonspecific immunoprecipitation condition (IgG). Error bars represent±standard error of three biological replicates

Plant Mol Biol Rep (2015) 33:624–637 631

containing complex at the GRXC9 promoter and inhibitsSA-mediated induction of the gene (Fig. 6). This result,together with previous evidence showing that the GRXC9protein can interact with TGA2 in the nucleus(Ndamukong et al. 2007), suggests that GRXC9 nega-tively controls its own gene through binding to TGAfactors while they are bound to the DNA, turning offgene expression.

Role of the SA-Dependent and NPR1-Independent Pathwayin the Stress Defense Response

The existence of an SA-dependent and NPR1-independentpathway for the expression of genes with a putative antioxi-dant and/or detoxifying roles in defense, such asGRXs, GSTs,and UGTs (coding for UDP-glucosyl transferases), has beenreported by our group and by others (Lieberherr et al. 2003;Blanco et al. 2005, 2009; Uquillas et al. 2004; Langlois-Meurinne et al. 2005; Fode et al. 2008). The induction oftwo GST (Lieberherr et al. 2003) and three UGT genes(Langlois-Meurinne et al. 2005), through an SA-dependentand NPR1-independent mechanism, was reported inArabidopsis plants inoculated with avirulent strains of Pseu-domonas syringae pv. tomato (Pst). Here, we validate theactivation of the GRXC9 gene by this pathway under anabiotic stress condition such as UVB radiation that, like theimmune reaction induced by avirulent Pst strains, triggers anSA-mediated defense response (Surplus et al. 1998).

On the other hand, members of GST and UGT genefamilies have also been found to be responsive totreatments with oxilipins (including JA) and xenobioticchemicals like 2,4-D, in the context of the chemicaldetoxification process. Based on the overlap of someGST and UGT (early NPR1-independent SAIG (Blancoet al. 2009) that are also responsive to oxilipins/xenobiotics (Fode et al. 2008; Baerson et al. 2005;Mueller et al. 2008), as well as on the involvementof common TGA factors and as-1-like elements in theirtranscriptional control (as discussed in the next sec-tion), it was assumed that exogenous treatments withSA unspecifically induced the chemical detoxificationprocess (Fode et al. 2008; Gatz 2012). Results shownin this work for GRXC9 and in other works for theGSTF2, GSTF6, UGT73B3, UGT73B5, and UGT73D1genes (Lieberherr et al. 2003; Langlois-Meurinne et al.2005) clearly argue against this idea, indicating thatthese antioxidant/detoxifying genes are activated byan endogenous stress-driven and SA-mediated pathway,which is distinct from the NPR1-dependent pathwaythat activates defense genes such as PR-1. The rapidand transient expression of genes with antioxidant anddetoxifying roles could be important to restrict theoxidative burst produced in the infected/damaged

tissues, avoiding the oxidative damage of systemictissues.

Mechanistic Aspects of the Transcriptional Control ofGRXC9Expression via an SA-Dependent and NPR1-IndependentPathway. Involvement of TGA Class II Factors and as-1-LikePromoter Elements

Evidence provided in this paper further supports the idea that,even though as-1-like and isolated TGACG boxes bind thesame class of TGA factors, they are functionally different.These factors respond to different pathways to control theexpression of distinct groups of genes that are activated atdifferent times during the defense response, using differentmechanism for promoter recognition and activation.

In Arabidopsis, as-1-like elements are overrepresented inpromoters of genes that code for enzymes with antioxidant ordetoxifying activity that is responsive to exogenous applica-tion of SA, xenobiotics, and oxilipins (Blanco et al. 2009;Fode et al. 2008; Köster et al. 2012). Functional requirementfor as-1-like elements has been previously reported for only acouple of these Arabidopsis genes: GST6 coding for a GSTinducible by xenobiotics, H2O2, and SA (Chen and Singh1999) and CYP81D11 coding for a cytochrome P450 induc-ible by xenobiotics (Köster et al. 2012). The functional anal-ysis of the as-1-like elements from the GRXC9 promoterdescribed in this work represents the first functional promoteranalysis performed in an early SA-dependent and NPR1-independent gene activated under an abiotic stress.

TheGRXC9 gene has two contiguous as-1-like elements inits promoter sequence (Fig. 2b and Online Resource 3). Inboth cases, the second TGACG box is less conserved than thefirst one, as previously described for other as-1-like elements(Ellis et al. 1993). The functional analysis of the GRXC9 genepromoter clearly indicates that both as-1-like elements areessential for SA-mediated expression of the gene (Fig. 2).Interestingly, the all or none effect of mutating any as-1-likeelement, instead of additive or synergistic effects, indicatesthat both elements must work together in the formation oftranscriptional complexes.

In contrast, isolated TGACG boxes are enriched in pro-moters of SAIGs by an NPR1-dependent pathway (Malecket al. 2000). Functional requirement of a TGACG box for SA-mediated expression was demonstrated for PR-1 and NIMIN1promoters, which are known to be induced by SA via anNPR1-dependent mechanism (Lebel et al. 1998; Fonsecaet al. 2010). In the case of the PR-1 gene, isolated TGACGboxes control its transcriptional activation by SA and itsrepression under basal conditions (Lebel et al. 1998).

Interestingly, this work shows that common TGA factors(TGA class II and TGA3) recognize as-1-like elements andTGACG boxes, being therefore involved in different tran-scriptional control processes. In Arabidopsis, TGA class II

632 Plant Mol Biol Rep (2015) 33:624–637

factors have been reported to be essential in several processes:the basal repression and the SA-mediated and NPR1-dependent induction of plant TGACG box-containing genes,such as PR1 gene (Rochon et al. 2006; Kesarwani et al. 2007;Zhang et al. 2003), the induction of as-1-like-containing plantgenes belonging to the chemical detoxification process, inresponse to treatments with xenobiotics and oxilipins (includ-ing JA) (Fode et al. 2008; Mueller et al. 2008; Stotz et al.2013), the induction of JA/ethylene-inducible genes likePDF1.2 and its negative modulation by SA (Ndamukonget al. 2007; Zander et al. 2010), and the SA-mediated induc-tion of early and NPR1-independent SAIGs (Blanco et al.2009), the last one supported by the results of this work.

One interesting conclusion of this work is that TGA2 isinvolved in different signaling pathways that operate at dif-ferent times in the response to stress (UVB in this case). Oneof these pathways leads to the expression of early SA-dependent and NPR1-independent genes, such as GRXC9;the other pathway leads to the expression of late SA- andNPR1-dependent genes, such as PR1 (Fig. 1). Furthermore, inthe case ofGRXC9, TGA2 is bound to the promoter during allthe phases of its expression profile (Fig. 4), being essential forthe transient increase in transcript levels (Fig. 3) and recruit-ment of the Pol II (Fig. 5). The question is that how is theactivity of TGA factors that are involved in different mecha-nisms of transcriptional control and that act at different timesafter stress controlled?

Results showed in this work prompt us to propose a model(Fig. 7) to explain how SA controls the transcription processof GRXC9 in the context of the defense response to stress.According to our model, TGA2 and TGA3 (forminghomodimers or heterodimers, T2/3) are constitutively boundto the two as-1-like elements of the GRXC9 promoter. Underbasal conditions, we propose that an inactive form of acoregulator complex (Co-RI) is bound to the TGA2-3/as-1-like complex, forming a transcriptionally inactive complex.Upon a stress condition, SA levels increase producing theactivation of the coregulator complex (switch from Co-RI toCo-RA). According to our results, Co-RA must provide thetransactivation activity for recruitment of the Pol II basalmachinery (Pol II complex) to initiate transcription. Accord-ing to our results with theGRXC9 overexpressor lines (Fig. 6),we propose that the GRXC9 protein is involved in turning offthe SA-mediated activation of its own gene through a directinteraction with the TGA2-3/as-1-like complex. So, onceGRXC9 gene expression is induced, the GRXC9 protein pro-duced and translocated to the nucleus could bind the TGA2-3/as-1-like complex, as indicated by Ndamukong et al. (2007).We propose that GRXC9 bound to the complex promotes theinactivation of the coregulator complex (Co-RA to Co-RI),switching from a transcriptionally active to a transcriptionallyinactive complex. GRXC9 expression is turned off becauseCo-RI does not have the ability to recruit Pol II basal

machinery. This mechanism would allow a rapid transcrip-tional response to stress signals, through transient changes inthe activity of a coregulator complex bound to the TGA2-TGA3/as-1-like platform complex preformed at the GRXC9promoter.

Results indicating that GRXC9 and TGA2 interact in thenucleus and that GRXC9 overexpression reduces the expres-sion of the 2,4-D-inducible CaMVas-1::GUS transgene andthe JA-inducible PDF1.2 gene (Ndamukong et al. 2007)support the idea that GRXC9 could play a more general rolein controlling the expression of TGA class II-target genes.Interestingly, the binding of other CC-typeGRX (ROXY1 andROXY2) to TGA9/10 factors and their role in anthers devel-opment (Murmu et al. 2010) supports a role for CC-typeGRXsin the control of gene expression. Considering that ROXY1must be expressed in the nucleus to complement roxy1mutantand that GRXC9 can also complement the roxy1 mutant (Liet al. 2009), it can be inferred that nuclear GRXC9 is requiredto exert its activity. Our results indicating thatGRXC9 binds tothe TGA2-3/as-1-like complex support this idea.

One of the key questions raised by this model is what is thenature of the proteins that form the coregulator complex,either in its active (Co-RA) or inactive (Co-RI) forms. Theexistence of a corepressor complex that binds the TGA2-3/as-1-like complex under basal conditions (Fig. 7) was previouslyproposed for the 35SCaMV as-1 in tobacco (Johnson et al.2001; Butterbrodt et al. 2006). In the case ofGRXC9, previousevidence showing that treatment with the protein inhibitorcycloheximide highly increases the basal and SA-inducedlevels of GRXC9 transcripts suggests the existence of a re-pressor with a high turnover rate (Blanco et al. 2009). Morerecently, a subunit of the mediator complex (MED18) wasfound to be essential for suppression of GRXC9 as well asGRXS13 and TRX-h5 expression in the absence of stress,suggesting that this subunit forms part of the Co-RI complex(Lai et al. 2014). On the other hand, we show here evidencethat although TGA2 and TGA3 are basally bound to theGRXC9 promoter, their binding is not required for turningoff the gene in the absence of stress. In fact, knockout mutantsfor TGA class II and TGA3 factors do not show increasedbasal levels of GRXC9 transcripts (Fig. 3). Together, thisevidence indicates that even though TGA2/TGA3 and Co-RI

(probably containing MED18) form part of the basal inactivecomplex, the presence of MED18 but not of TGA2/TGA3factors is essential to repress GRXC9 transcription.

Concerning proteins with coactivator function that could bepart of the Co-RA complex, we discard NPR1 and SCL14proteins. In fact, the induction of GRXC9 by SA is not onlyindependent of the NPR1 protein ((Blanco et al. 2009) andOnline Resource 1) but also independent of the SCL14 protein(Fode et al. 2008), which was previously identified as aninteractor of TGA2 factor essential for the expression of agroup of as-1-like-containing genes. Previously, a Dof protein

Plant Mol Biol Rep (2015) 33:624–637 633

named OBP1 was found to interact with TGA4 and TGA5 toenhance binding to the as-1 element (Zhang et al. 1995);whether this kind of protein forms part of the Co-RA complexremains to be elucidated. Therefore, further efforts are re-quired for the identification of the protein(s) that form partof the Co-RA complex that binds to the TGA2-3/as-1-likecomplex in GRXC9 promoter.

Another interesting question raised by this model iswhat is the mechanism by which SA promotes theactivation of the coregulator complex (switch fromCo-RI to Co-RA), as well as how GRXC9 can promotethe inactivation of the coregulator complex (switchfrom Co-RA to Co-RI). We speculate in our model thata redox change promoted by SA accumulation can beresponsible for the activation of the coregulator com-plex. This idea is supported by evidence indicating thatSA promotes a biphasic change in the GSH/GSSGratio, first an oxidative phase characterized by a de-crease in GSH/GSSG ratio and then a reductive phasecharacterized by increase in GSH/GSSG ratio (Mouet al. 2003; Mateo et al. 2006). With respect to themechanism of inactivation of the coregulator complex,we speculate that GRXC9, through its oxidoreductaseactivity, can catalyze the reduction of a protein thatforms part of the coregulator complex, producing its

inactivation. GRXC9 could be a key piece in the redoxcontrol of the expression of genes controlled by TGAclass II factors.

Methods

Plant Growth Conditions and Treatments

Arabidopsis thaliana wild-type (WT), npr1-1 (Cao et al.1994), tga1-1/tga4-1, tga3-1, tga2-1/tga3-1/tga5-1/tga6-1(Kesarwani et al. 2007) tga-7-1, tga2-1/tga5-1/tga6-1 (Zhanget al. 2003), and sid2-2 (Wildermuth et al. 2001) plants were inColumbia (Col-0) background. Seedlings were grown in vitroin 0.5× MS medium supplemented with 10 g/l sucrose and2.6 g/l Phytagel (Sigma) under controlled conditions (16 hlight, 80 μmol/m2/s, 22±2 °C). For ChIP and gene expressionassays, 15-day-old seedlings were floated on 0.5 mM SA(treatment) or 0.5× MS medium as a control and incubatedfor the indicated periods of time under continuous light(80 μmol/m2/s). For gene expression assays, whole seedlingswere immediately frozen in liquid nitrogen and stored at−70 °C until RNA isolation. For ChIP assays, whole seedlingswere processed immediately as described below. For UVBirradiation assays, 15-day-old seedlings were exposed to UVB

Fig. 7 Mechanistic model for the transcriptional control of GRXC9expression by stress, via an SA-dependent and NPR1-independent path-way in Arabidopsis. Homodimers or heterodimers of TGA2 and TGA3(T2/3) are constitutively bound to the two as-1-like elements of theGRXC9 promoter, acting as a platform for the formation of transcription-ally inactive and active complexes. Under basal conditions, an inactiveform of a coregulator (Co-RI) is bound to the TGA2-3/as-1-like complexforming a basal complex that impairs recruitment of the Pol II to theGRXC9 promoter. Upon stress, SA is rapidly accumulated promoting theactivation of the coregulator complex (switch from Co-RI to Co-RA) that

binds to the TGA2-3/as-1-like complex, allowing the formation of atranscriptionally active complex that recruits the Pol II basal machinery(Pol II complex) to the GRXC9 basal promoter. Transcription of GRXC9leads to the accumulation of the GRXC9 protein in the nucleus where itbinds to the TGA2-3/as-1-like complex producing the inactivation of thecoregulator complex (switch from Co-RA to Co-RI) and therefore turningoffGRXC9 transcription. We speculate that the switch from Co-RI to Co-RA promoted by SA is produced by the oxidative modification of one ofthe proteins involved in the promoter complex, while the switch from Co-RA to Co-RI is produced by the protein’s reduction catalyzed by GRXC9

634 Plant Mol Biol Rep (2015) 33:624–637

light (0.07 mW/cm2) in a chamber equipped with two USHIOUVB F8T5.UB-V, UVP 3400401 fluorescent tubes (λ=306 nm). As a control, we used nonirradiated seedlings.

Genetic Constructs and Plant Transformation

Genetic constructs were generated using the Gatewaytechnology following the manufacturer’s instructions(Invitrogen). The GRXC9 promoter regions including the5′UTR, −1,849 to +26; −168 to +26; −112 to +26; and−61 to +26, called in the text as pC9 WT, pC9-168, pC9-112, and pC9-61, respectively, were obtained by amplifi-cation from genomic DNA, using the oligonucleotidesindicated in Online Resource 8. PCR fragments werecloned into the pENTR/SD/D-TOPO vector and thenrecombined into the pKGWFS7 vector to generate tran-scriptional fusions with eGFP and β-glucuronidase GUSreporter genes (Karimi et al. 2002). Site-directed mutationof the distal and the proximal as-1-like element wasperformed on the pC9 WT promoter fragment cloned intothe pENTR/SD/TOPO vector, as previously described(Weiner et al. 1994). The site-directed mutations generat-ed are shown in Fig. 2c, and the oligonucleotides used toproduce these mutations are listed in the Online Resource8. The purified PCR products were recombined into thepKGWFS7 vector. In order to generate GRXC9overexpressor lines, the GRXC9-coding region was am-plified from complementary DNA (cDNA) using theprimers described in the Online Resource 8. The PCRproduct was cloned into the pENTR/SD/D-TOPO vectorand then recombined into the pBADcMyc vector to ex-press the GRXC9 protein fused to a c-Myc tag controlledby the 35S CaMV promoter. Final constructs were veri-fied by sequencing and introduced into the Agrobacteriumtumefaciens C58 strain. Arabidopsis plants were trans-formed by floral dip method. Transgenic seeds were se-lected in 0.5× MS solid medium supplemented with50 μg/ml kanamycin for the GUS reporter lines or15 μg/ml glufosinate-ammonium for the overexpressorlines. Stable homozygous transgenic lines were used forfurther analyses.

GUS Assays

GUS activity was determined in control- and SA-treated seed-lings from each transgenic line carrying theGRXC9 promoter-driven GUS constructs described above. The 4-methylumbelliferyl-D-glucuronide was used as substrate, andthe fluorescent product 4-methylumbelliferone was quanti-fied, as previously described (Jefferson et al. 1987). Treat-ments were done by triplicate for each line, and the measure-ments were normalized with total protein content quantifiedusing the Bradford assay (Bio-Rad).

ChIPAssays

ChIP assays were performed as described (Saleh et al. 2008).Five microliters of the following antibodies were used for im-munoprecipitation assays: Pol II polyclonal antibody (sc-33754,Santa Cruz Biotechnology), TGA1, TGA2, and TGA3 polyclon-al antibodies (Lam and Lam 1995), c-Myc polyclonal antibody(A-14, sc-789, Santa Cruz), and normal purified IgG (A2609,Santa Cruz Biotechnology) used as control of a nonspecificantibody. The concentration of DNA in each sample (inputchromatin and chromatin immunoprecipitated with either specif-ic or nonspecific antibodies) was quantified by qPCR, using theStratageneMX3000P® equipment and the Sensimix Plus SYBRGreen Reagents (Quantece). Primers used to amplify theGRXC9promoter region containing the as-1-like elements (−212 to +78)are listed in Online Resource 8.

Gene Expression Analysis

Total RNA was obtained from frozen samples using theTRIzol® Reagent (Invitrogen) according to the manufac-turer’s instructions. cDNAwas synthesized from each sample(2 μg of total RNA) with an ImProm II Kit (Promega). qPCRwas performed with the Stratagene MX3000P® equipment.The expression levels of GRXC9 and PR-1 were calculatedrelative to the YLS8 (AT5G08290) or Clathrin adaptor com-plex subunit (AT4G24550) genes. Primers used for each geneare listed in Online Resource 8.

Yeast Two-Hybrid Assays

The coding regions of TGA factors (TGA2, TGA5, and TGA6and TGA3) were cloned into the pDONR201 vector (Jakobyet al. 2002). These coding regions were then recombined intothe pDEST22 vector, to produce a fusion protein with theGal4 DNA-binding domain and into the pDEST32 vector, toproduce a fusion protein with the transactivation domain ofthe Gal4 factor. Different combinations of two constructs wereused to transform the SFY526 yeast strain (harboring theGal4RE::β-Gal reporter construct), and qualitative assays forβ-Gal activity were performed as described (Gietz andSchiestl 2007). Interaction between NPR1 and TGA2 wasassayed as a positive control, and a combination of thepDEST32 and pDEST22 empty vectors was used as a nega-tive control.

Acknowledgments The authors thank Dr. Eric Lam (Rutgers, TheState University of New Jersey, School of Environmental and BiologicalSciences, Department of Plant Biology and Pathology) for providingantibodies against TGA factors.

Funding This work was supported by the National Commission forScience and Technology CONICYT (FONDECYT grant nos. 1100656

Plant Mol Biol Rep (2015) 33:624–637 635

and 1141202), the Millennium Science Initiative (Nucleus for PlantFunctional Genomics, grant no. P10-062-F), and the Spanish Ministryof Science and Innovation (project number: BIO2010-14871). A.H. wassupported by a PhD fellowship from CONICYT.

Open Access This article is distributed under the terms of the CreativeCommons Attribution License which permits any use, distribution, andreproduction in any medium, provided the original author(s) and thesource are credited.

References

Baerson SR, Sánchez-Moreiras A, Pedrol-Bonjoch N, Schulz M, KaganIA, Agarwal AK, Reigosa MJ, Duke SO (2005) Detoxification andtranscriptome response in Arabidopsis seedlings exposed to theallelochemical benzoxazolin-2(3H)-one. J Biol Chem 280(23):21867–21881. doi:10.1074/jbc.M500694200

Benfey P, Ren L, Chua N (1989) The CaMV 35S enhancer contains atleast two domains which can confer different developmental andtissue-specifi expression patterns. EMBO J 8(8):2195–2202

Blanco F, Garreton V, Frey N, Dominguez C, Perez-Acle T, Van derStraeten D, Jordana X, Holuigue L (2005) Identification of NPR1-dependent and independent genes early induced by salicylic acidtreatment in Arabidopsis. Plant Mol Biol 59(6):927–944. doi:10.1007/s11103-005-2227-x

Blanco F, Salinas P, Cecchini NM, Jordana X, Van Hummelen P, AlvarezME, Holuigue L (2009) Early genomic responses to salicylic acid inArabidopsis. Plant Mol Biol 70(1–2):79–102. doi:10.1007/s11103-009-9458-1

Borsani O, Valpuesta V, Botella MA (2001) Evidence for a role ofsalicylic acid in the oxidative damage generated by NaCl andosmotic stress in Arabidopsis seedlings. Plant Physiol 126(3):1024–1030

Butterbrodt T, Thurow C, Gatz C (2006) Chromatin immunoprecipitationanalysis of the tobacco PR-1a- and the truncated CaMV 35S pro-moter reveals differences in salicylic acid-dependent TGA factorbinding and histone acetylation. Plant Mol Biol 61(4–5):665–674.doi:10.1007/s11103-006-0039-2

Cao H, Bowling SA, Gordon SA, Dong X (1994) Characterization of anArabidopsis mutant that is nonresponsive to inducers of systemicacquired resistance. Plant Cell 6(11):1583–1592. doi:10.1105/tpc.6.11.1583

Chen W, Singh KB (1999) The auxin, hydrogen peroxide and salicylicacid induced expression of the Arabidopsis GST6 promoter ismediated in part by an ocs element. Plant J 19(6):667–677. doi:10.1046/j.1365-313x.1999.00560.x

Dong X (2004) NPR1, all things considered. Curr Opin Plant Biol 7(5):547–552. doi:10.1016/j.pbi.2004.07.005

Dubreuil-Maurizi C, Poinssot B (2012) Role of glutathione in plantsignaling under biotic stress. Plant Signal Behav 7(2):210–212.doi:10.4161/psb.18831

Ellis JG, Tokuhisa JG, Llewellyn DJ, Bouchez D, Singh K, Dennis ES,Peacock WJ (1993) Does the ocs-element occur as a functionalcomponent of the promoters of plant genes? Plant J 4(3):433–443.doi:10.1046/j.1365-313X.1993.04030433.x

Fan W, Dong X (2002) In vivo interaction between NPR1 and transcrip-tion factor TGA2 leads to salicylic acid-mediated gene activation inArabidopsis. Plant Cell 14(6):1377–1389

Fode B, Siemsen T, ThurowC,Weigel R, Gatz C (2008) TheArabidopsisGRAS protein SCL14 interacts with class II TGA transcription

factors and is essential for the activation of stress-inducible pro-moters. Plant Cell 20(11):3122–3135. doi:10.1105/tpc.108.058974

Fonseca J, Menossi M, Thibaid-Nissen F, Town C (2010) Functionalanalysis of a TGA factor-binding site located in the promoter regioncontrolling salicylic acid-induced NIMIN-1 expression inArabidopsis. Genet Mol Res 9(1):167–175

Foyer CH, Noctor G (2011) Ascorbate and glutathione: the heart of theredox hub. Plant Physiol 155(1):2–18. doi:10.1104/pp. 110.167569

Fu ZQ, Dong X (2013) Systemic acquired resistance: turning localinfection into global defense. Annu Rev Plant Biol 64:839–863.doi:10.1146/annurev-arplant-042811-105606

Fu ZQ, Yan S, Saleh A, Wang W, Ruble J, Oka N, Mohan R, Spoel SH,Tada Y, Zheng N, Dong X (2012) NPR3 and NPR4 are receptors forthe immune signal salicylic acid in plants. Nature 486(7402):228–232. doi:10.1038/nature11162

Garcion C, LohmannA, Lamodiere E, Catinot J, Buchala A, Doermann P,Metraux JP (2008) Characterization and biological function of theISOCHORISMATE SYNTHASE2 gene of Arabidopsis. PlantPhysiol 147(3):1279–1287. doi:10.1104/pp. 108.119420

Garreton V, Carpinelli J, Jordana X, Holuigue L (2002) The as-1 promot-er element is an oxidative stress-responsive element and salicylicacid activates it via oxidative species. Plant Physiol 130(3):1516–1526. doi:10.1104/pp. 009886

Gatz C (2012) From pioneers to team players: TGA transcription factorsprovide a molecular link between different stress pathways. Mol PlantMicrobe Interact 26(2):151–159. doi:10.1094/mpmi-04-12-0078-ia

Gietz D, Schiestl RH (2007) High-efficiency yeast transformation usingthe LiAc/SS carrier DNA/PEG method. Nat Protocols 2(1):31–34

Han Y, Chaouch S, Mhamdi A, Queval G, Zechmann B, Noctor G (2013)Functional analysis of Arabidopsismutants points to novel roles forglutathione in coupling H(2)O(2) to activation of salicylic acidaccumulation and signaling. Antioxid Redox Signal 18(16):2106–2121. doi:10.1089/ars.2012.5052

Jakoby M, Weisshaar B, Droge-Laser W, Vicente-Carbajosa J,Tiedemann J, Kroj T, Parcy F (2002) bZIP transcription factors inArabidopsis. Trends Plant Sci 7(3):106–111

Jefferson R, Kavanagh T, Bevan M (1987) GUS fusions: beta-glucuronidase as a sensitive and versatile gene fusion marker inhigher plants. EMBO J 6(13):3901–3907

Johnson C, Glover G, Arias J (2001) Regulation of DNA binding andtrans-activation by a xenobiotic stress-activated plant transcriptionfactor. J Biol Chem 276(1):172–178. doi:10.1074/jbc.M005143200

Karimi M, Inzé D, Depicker A (2002) GATEWAY™ vectors forAgrobacterium-mediated plant transformation. Trends Plant Sci7(5):193–195. doi:10.1016/S1360-1385(02)02251-3

Kesarwani M, Yoo J, Dong X (2007) Genetic interactions of TGAtranscription factors in the regulation of pathogenesis-related genesand disease resistance in Arabidopsis. Plant Physiol 144(1):336–346. doi:10.1104/pp. 106.095299

Köster J, Thurow C, Kruse K, Meier A, Iven T, Feussner I, Gatz C (2012)Xenobiotic- and jasmonic acid-inducible signal transduction pathwayshave become interdependent at the ArabidopsisCYP81D11 promoter.Plant Physiol 159(1):391–402. doi:10.1104/pp. 112.194274

Krawczyk S, Thurow C, Niggeweg R, Gatz C (2002) Analysis of thespacing between the two palindromes of activation sequence-1 withrespect to binding to different TGA factors and transcriptionalactivation potential. Nucleic Acids Res 30(3):775–781

La Camera S, L’Haridon F, Astier J, Zander M, Abou-Mansour E, PageG, Thurow C, Wendehenne D, Gatz C, Metraux JP, Lamotte O(2011) The glutaredoxin ATGRXS13 is required to facilitateBotrytis cinerea infection of Arabidopsis thaliana plants. Plant J68(3):507–519. doi:10.1111/j.1365-313X.2011.04706.x

Lai Z, Schluttenhofer C, Bhide K, Shreve J, Thimmapuram J, Lee S, YunD-J, Mengiste T (2014) MED18 interaction with distinct transcrip-tion factors regulates multiple plant functions. Nat Commun 5. doi:10.1038/ncomms4064|www.nature.com/naturecommunications

636 Plant Mol Biol Rep (2015) 33:624–637

Lam E, Lam YK-P (1995) Binding site requirements and differentialrepresentation of TGA factors in nuclear ASF-1 activity. NucleicAcids Res 23(18):3778–3785. doi:10.1093/nar/23.18.3778

Langlois-Meurinne M, Gachon CMM, Saindrenan P (2005) Pathogen-responsive expression of glycosyltransferase genes UGT73B3 andUGT73B5 is necessary for resistance to pseudomonas syringae pvtomato in Arabidopsis. Plant Physiol 139(4):1890–1901. doi:10.1104/pp. 105.067223

Laporte D, Olate E, Salinas P, Salazar M, Jordana X, Holuigue L (2012)Glutaredoxin GRXS13 plays a key role in protection against pho-tooxidative stress in Arabidopsis. J Exp Bot 63(1):503–515. doi:10.1093/jxb/err301

Lebel E, Heifetz P, Thorne L, Uknes S, Ryals J,Ward E (1998) Functionalanalysis of regulatory sequences controlling PR-1 gene expressionin Arabidopsis. Plant J 16(2):223–233. doi:10.1046/j.1365-313x.1998.00288.x

Lee S, Kim SG, Park CM (2010) Salicylic acid promotes seed germina-tion under high salinity by modulating antioxidant activity inArabidopsis. New Phytol 188(2):626–637. doi:10.1111/j.1469-8137.2010.03378.x

Li S, Lauri A, ZiemannM, Busch A, Bhave M, Zachgo S (2009) Nuclearactivity of ROXY1, a glutaredoxin interacting with TGA factors, isrequired for petal development in Arabidopsis thaliana. Plant Cell21(2):429–441. doi:10.1105/tpc.108.064477

Lieberherr D, Wagner U, Dubuis P-H, Métraux J-P, Mauch F (2003) Therapid induction of glutathione S-transferases AtGSTF2 andAtGSTF6 by avirulent pseudomonas syringae is the result of com-bined salicylic acid and ethylene signaling. Plant Cell Physiol 44(7):750–757. doi:10.1093/pcp/pcg093

MaleckK, Levine A, EulgemT,MorganA, Schmid J, Lawton KA, DanglJL, Dietrich RA (2000) The transcriptome of Arabidopsis thalianaduring systemic acquired resistance. Nat Genet 26(4):403–410

Mateo A, Funck D,Muhlenbock P, Kular B,Mullineaux PM,Karpinski S(2006) Controlled levels of salicylic acid are required for optimalphotosynthesis and redox homeostasis. J Exp Bot 57(8):1795–1807.doi:10.1093/jxb/erj196

Miura K, Okamoto H, Okuma E, Shiba H, Kamada H, Hasegawa PM,Murata Y (2013) SIZ1 deficiency causes reduced stomatal apertureand enhanced drought tolerance via controlling salicylic acid-induced accumulation of reactive oxygen species in Arabidopsis.Plant J 73:91–104. doi:10.1111/tpj.12014

Mou Z, Fan W, Dong X (2003) Inducers of plant systemic acquiredresistance regulate NPR1 function through redox changes. Cell113(7):935–944. doi:10.1016/S0092-8674(03)00429-x

Mueller S, Hilbert B, Dueckershoff K, Roitsch T, Krischke M, MuellerMJ, Berger S (2008) General detoxification and stress responses aremediated by oxidized lipids through TGA transcription factors inArabidopsis. Plant Cell 20(3):768–785. doi:10.1105/tpc.107.054809

Murmu J, Bush MJ, DeLong C, Li S, Xu M, Khan M, Malcolmson C,Fobert PR, Zachgo S, Hepworth SR (2010) Arabidopsis basicleucine-zipper transcription factors TGA9 and TGA10 interact withfloral glutaredoxins ROXY1 and ROXY2 and are redundantlyrequired for anther development. Plant Physiol 154(3):1492–1504.doi:10.1104/pp. 110.159111

Nawrath C, Metraux JP (1999) Salicylic acid induction-deficient mutantsof Arabidopsis express PR-2 and PR-5 and accumulate high levelsof camalexin after pathogen inoculation. Plant Cell 11(8):1393–1404

Ndamukong I, Abdallat AA, Thurow C, Fode B, Zander M, Weigel R,Gatz C (2007) SA-inducibleArabidopsis glutaredoxin interacts withTGA factors and suppresses JA-responsive PDF1.2 transcription.Plant J 50(1):128–139. doi:10.1111/j.1365-313X.2007.03039.x

Noshi M, Maruta T, Shigeoka S (2012) Relationship between chloroplas-tic H 2O 2 and the salicylic acid response. Plant Signal Behav 7(8):944–946. doi:10.4161/psb.20906

OgawaD, NakajimaN, Sano T, TamaokiM, AonoM,KuboA,KannaM,Ioki M, Kamada H, Saji H (2005) Salicylic acid accumulation underO3 exposure is regulated by ethylene in tobacco plants. Plant CellPhysiol 46(7):1062–1072. doi:10.1093/pcp/pci118

Qin XF, Holuigue L, Horvath DM, Chua NH (1994) Immediate earlytranscription activation by salicylic acid via the cauliflower mosaicvirus as-1 element. Plant Cell 6(6):863–874. doi:10.1105/tpc.6.6.863

Rochon A, Boyle P, Wignes T, Fobert PR, Després C (2006) Thecoactivator function of Arabidopsis NPR1 requires the core of itsBTB/POZ domain and the oxidation of C-terminal cysteines. PlantCell 18(12):3670–3685. doi:10.1105/tpc.106.046953

Rouhier N, Lemaire SD, Jacquot JP (2008) The role of glutathione inphotosynthetic organisms: emerging functions for glutaredoxins andglutathionylation. Annu Rev Plant Biol 59:143–166. doi:10.1146/annurev.arplant.59.032607.092811

Saleh A, Alvarez-Venegas R, Avramova Z (2008) An efficient chromatinimmunoprecipitation (ChIP) protocol for studying histone modifi-cations in Arabidopsis plants. Nat Protocols 3(6):1018–1025. doi:10.1038/nprot.2008.66

Shearer HL, ChengYT,Wang L, Liu J, Boyle P, Després C, ZhangY, Li X,Fobert PR (2012) Arabidopsis clade I TGA transcription factorsregulate plant defenses in an NPR1-independent fashion. Mol PlantMicrobe Interact 25(11):1459–1468. doi:10.1094/mpmi-09-11-0256

Stotz HU, Mueller S, Zoeller M, Mueller MJ, Berger S (2013) TGAtranscription factors and jasmonate-independent COI1 signallingregulate specific plant responses to reactive oxylipins. J Exp Bot64(4):963–975. doi:10.1093/jxb/ers389

Surplus SL, Jordan BR, Murphy AM, Carr JP, Thomas B, MackernessSAH (1998) Ultraviolet-B-induced responses in Arabidopsisthaliana: role of salicylic acid and reactive oxygen species in theregulation of transcripts encoding photosynthetic and acidicpathogenesis-related proteins. Plant Cell Environ 21(7):685–694.doi:10.1046/j.1365-3040.1998.00325.x

Tada Y, Spoel SH, Pajerowska-Mukhtar K, Mou Z, Song J, Wang C, ZuoJ, Dong X (2008) Plant immunity requires conformational changes[corrected] of NPR1 via S-nitrosylation and thioredoxins. Science321(5891):952–956. doi:10.1126/science.1156970

Uquillas C, Letelier I, Blanco F, Jordana X, Holuigue L (2004) NPR1-independent activation of immediate early salicylic acid-responsivegenes in Arabidopsis. Mol Plant Mibrobe Interact 17(1):34–42. doi:10.1094/MPMI.2004.17.1.34

Vlot AC, Dempsey DA, Klessig DF (2009) Salicylic acid, a multifacetedhormone to combat disease. Annu Rev Phytopathol 47:177–206.doi:10.1146/annurev.phyto.050908.135202

Weiner MP, Costa GL, Schoettlin W, Cline J, Mathur E, Bauer JC (1994)Site-directed mutagenesis of double-stranded DNA by the polymer-ase chain reaction. Gene 151(1–2):119–123. doi:10.1016/0378-1119(94)90641-6

Wildermuth MC, Dewdney J, Wu G, Ausubel FM (2001) Isochorismatesynthase is required to synthesize salicylic acid for plant defence.Nature 414(6863):562–565. doi:10.1038/35107108

Zander M, La Camera S, Lamotte O, Métraux J-P, Gatz C (2010)Arabidopsis thaliana class-II TGA transcription factors are essentialactivators of jasmonic acid/ethylene-induced defense responses.Plant J 61(2):200–210. doi:10.1111/j.1365-313X.2009.04044.x

Zhang B, Chen W, Foley R, Büttner M, Singh K (1995) Interactionsbetween distinct types of DNA binding proteins enhance binding toocs element promoter sequences. Plant Cell 7(12):2241–2252. doi:10.1105/tpc.7.12.2241

Zhang Y, Tessaro MJ, Lassner M, Li X (2003) Knockout analysis ofArabidopsis transcription factors TGA2, TGA5, and TGA6 revealstheir redundant and essential roles in systemic acquired resistance.Plant Cell 15(11):2647–2653. doi:10.1093/jxb/ers389

ZhangY,ChengYT,QuN, ZhaoQ,BiD, LiX (2006)Negative regulation ofdefense responses in Arabidopsis by two NPR1 paralogs. Plant J 48(5):647–656. doi:10.1111/j.1365-313X.2006.02903.x

Plant Mol Biol Rep (2015) 33:624–637 637